AI Created images (generated by Midjourney) lose registered copyrights

[05032-Mendicant-Bias]

10 minutes ago, goodtofufriday said:

So why is someone who tells a machine "make this piece of art" an artist? 

It takes more steps than you think to generate a decent looking image.

 

In this example I used OpenPose to get the pose from the left image, used ProtogenX34 guided by open pose to generate a 1024x1024 blurry mess with mishaped face and fingers, used GAN4X to upscale with a GAN network, then inpainted and regenerated individual details like hands, face, etc... All while refining positive and negative prompts.

It's not art skills because I can never draw anything even remotely close to the right image, no matter how much time I would choose to learn it. It is certanly creative skills to refine the prompt and get closer to what I had in mind. A D&D high elf mage character.

[Holmes108]

Also, while I just wrote a post defending the concept of granting copyright to AI, I'm not necessarily ready to die on that hill, or claim I can't be convinced otherwise. For example, an argument on the other side:

 

Not sure if this has been brought up yet, but what about the danger of AI becoming the equivalent of a copyright/patent troll. What if I can make a computer spit out literally millions (maybe billions?) of pictures in the span of a day. Even if they pass the "originality" test for us all... they could perhaps come too close to owning all too many variations of expression, breaking the system...

 

The world is about to get very messy.

[Holmes108]

14 minutes ago, Kinda Bottlenecked said:

Seems like a benefit to society if AI art isn't copyrightable. Which means it can't be taken down or claimed either I assume.

That's where I get lost. If it's not copyrightable, then does that mean someone else can recreate it and copyright it? That would seem broken. So then instead it becomes some sort of public domain then, I suppose? Which still implies some sort of ownership, in a sense, but you're just not entitled to exclusive use of it as the AI (or AI owner)?

 

ugh, lol. We aren't ready for this.

[graograman]

4 minutes ago, Kinda Bottlenecked said:

Seems like a benefit to society if AI art isn't copyrightable. Which means it can't be taken down or claimed either I assume.

Shutterstock and friends are already rubbing their greedy hands. They will absolutely claim and take down AI-generated Art for easy money. Making content using AI-generated Art could become a huge liability.

[Zodiark1593]

2 hours ago, jaslion said:

Thing is it's not because it has repeatedly committed plagiarism over and over including showing signatures of others

If artists really want legislation to go one step further, it would be possible to exclude non-humans from Fair Use protection, in addition to copyright protection. This would pretty much eliminate the argument over the ethics of using material without consent, as it would no longer matter. As for who takes the liability, and how enforceable this would be (if training data was created and distributed anonymously, finding a liable party may be impossible), is certainly up for debate. 
 

43 minutes ago, goodtofufriday said:

My opnion overall summed up by Miyazaki when AI generated art/animation was pitched to him.Im a IT admin in corporate world and I see nothing good coming from AI for civilzation in general. Humans have had a hard enough time adapting socially to the explosion in technology in the past 30 years. We arent ready for AI. 

 

As of now, it’s entirely impossible to restrict or regulate what can be developed and ran on PC hardware. The resources and knowledge to develop AI is freely available, regardless of one’s moral compass. 
 

The reality is that, whether humanity is ready or not, is irrelevant. There is no government body in a position to do anything about it. 

[Kinda Bottlenecked]

30 minutes ago, 05032-Mendicant-Bias said:

It takes more steps than you think to generate a decent looking image.

 

In this example I used OpenPose to get the pose from the left image, used ProtogenX34 guided by open pose to generate a 1024x1024 blurry mess with mishaped face and fingers, used GAN4X to upscale with a GAN network, then inpainted and regenerated individual details like hands, face, etc... All while refining positive and negative prompts.

It's not art skills because I can never draw anything even remotely close to the right image, no matter how much time I would choose to learn it. It is certanly creative skills to refine the prompt and get closer to what I had in mind. A D&D high elf mage character.

 

what do you think made the image you generated in your post look good?

[BabaGanuche]

4 minutes ago, 05032-Mendicant-Bias said:

It's not art skills because I can never draw anything even remotely close to the right image, no matter how much time I would choose to learn it. It is certanly creative skills to refine the prompt and get closer to what I had in mind. A D&D high elf mage character.

You decisions in combing the AI tools may be creative enough to qualify you as the artist but that is something that I am not sure about since most of the output was not create by you. Something like how phone number are not copyrightable but the arrangement of can be copyrightable. I would think there is some amount work

 

My concern over all of this as I see AI art as a very cool and beneficial to society tool assuming that is handled properly. It would allow more people to have art for things that would otherwise would not. Your example is a perfect one for why I think it can be a good thing.

 

This is one of my main concerns on how this can turn into something very bad.

16 minutes ago, Holmes108 said:

Not sure if this has been brought up yet, but what about the danger of AI becoming the equivalent of a copyright/patent troll. What if I can make a computer spit out literally millions (maybe billions?) of pictures in the span of a day. Even if they pass the "originality" test for us all... they could perhaps come too close to owning all too many variations of expression, breaking the system...

 

11 minutes ago, Holmes108 said:

That's where I get lost. If it's not copyrightable, then does that mean someone else can recreate it and copyright it?

No, as it would not be original.

 

https://guides.lib.umich.edu/copyrightbasics/copyrightability

 

[05032-Mendicant-Bias]

15 minutes ago, Kinda Bottlenecked said:

what do you think made the image you generated in your post look good?

Simple, I like it! The bulk of the creativity was inferred from this database (https://laion.ai/blog/laion-400-open-dataset/)

It's fair value? 0$.
 

I like the AI generating process a lot more than searching for the closest image to what I want on google search. E.g. it was hard getting the crown thingie on her head. I wanted to get a side of her face scarred, but I couldn't do it. The model doesn't know scars. I'll have fun finding ways to do it.

 

10 minutes ago, BabaGanuche said:

My concern over all of this as I see AI art as a very cool and beneficial to society tool assuming that is handled properly. It would allow more people to have art for things that would otherwise would not. Your example is a perfect one for why I think it can be a good thing.

I like to think non programmers are being empowered the same way by chatgpt, as I'm empowered to generate image as a non artist. 

 

I think AI Generators are our civilization's ticket out of a lot of problems endemic to our society. We all need to become better at telling counterfeit and fake news, and that I believe will have positive rippling effect. E.g. it's about time education movedaway from testing thing GPT can give you an answer with one query.

[wanderingfool2]

3 hours ago, LAwLz said:

Why is it good that AI generated art can't be copyrighted? 

Things are going to get messy very quickly.

 

28 minutes ago, Kinda Bottlenecked said:

Seems like a benefit to society if AI art isn't copyrightable. Which means it can't be taken down or claimed either I assume.

It depends I think.  The issue about having "AI art" non copyrightable is that it can get very murky in terms of where the line is drawn and it can take away some of the rights people might expect.  Specifically around how much user interaction should be required for something to be copyrightable.

 

As an example, if they trained it only on images they drew should it be copyrightable? (I'd argue yes)

Now what happens if lets say you provide a skeleton frame that you wish to animate and you use AI and a reference image to do it?

What if you slowly describe and inpaint an image to get an image you want?

What if you describe an entire scene and it makes that scene, is it specific enough?

 

The issue I have with AI art being non-copyrightable is that I do believe there are some things that should be copyrightable (maybe they would lose certain aspects of copyright, but I think it shouldn't be outright not copyrightable). 

[Thaldor]

38 minutes ago, goodtofufriday said:

I think your arguement generally fails because its humans interacting with humans with humans doing all the interpreating. 

If we had the food making machines from star trek, would you call someone who told a machine "make a michelin 5 star meal" a chef? Probably not. So why is someone who tells a machine "make this piece of art" an artist? 

It all about humans hands or mind physically interacting with the space around us, taking bits and pecies of things to create something new. A machine doing the bits and pieces part is not the human creating. 

And I still question, how is it any different if human tells human to do piece of art and goes around saying they made the piece?

 

As in how it is any different if Gordon Ramsay tells his kitchen army to make a sandwich and then goes to the public saying he made the sandwich against Gordon Ramsay telling the fabricator to make him a sandwich and goes around saying he made the sandwich? In both cases I would say Gordon Ramsay didn't make the sandwich and I wouldn't call him the sandwich chef before he made the sandwich by himself with his own little/big hands.

 

Actually I think it is even worse what art world really does than what they try to fight against because the very point of copyrights is that the person who made the piece has rights to it, even when it is done in collaboration. Jeff Koons and whoever else who doesn't credit their studio assistants for the work they do are IMO way worse than an artist who claims copyright for a piece made by AI because artists not crediting their studio assistants really do steal the origin rights from persons who by the very copyright laws would have those rights instead of taking the rights from something that cannot even have the rights by the law as it isn't a person and so protected by the copyright laws.

 

But as the world is we are fine by Jeff Koons going and saying to one of his studio assistants to make Pantone color X colored glass sculpture of a balloon dog in his style and, vóla, there is a new Balloon dog sculpture by Jeff Koons. And the studio assistant got a paycheck and marking to their CV to be worked in Jeff Koons studio (and AFAIK they cannot refer to the pieces they made in their portfolio even). But most likely by the article if Jeff Koons was to tell AI controlled machine to make Pantone color X colored glass sculpture of a balloon dog in style of Jeff Koons, nope, not a new Balloon dog sculpture by Jeff Koons because he didn't steal the credits from a natural person.

[jagdtigger]

3 hours ago, LAwLz said:

Why is it good that AI generated art can't be copyrighted? 

Because its a generator and not a creator. It takes ppls work and use parts from them without any sort of modification and stitches them together......


As for the topic its a good decision. Now the only thing we need is a ban on using copyrighted material without permission as training material.

[jaslion]

15 minutes ago, Kinda Bottlenecked said:

 

what do you think made the image you generated in your post look good?

Skills of others on which the ai was trained.

 

I am quite upset as art of mine was absorbed into learnig (likely 100 pieces+) and I've fucked up looking versions of that signature a couple dozen times now which is upsetting to know that 7 years of work gets recreated in 7 seconds

[wanderingfool2]

9 minutes ago, Thaldor said:

And I still question, how is it any different if human tells human to do piece of art and goes around saying they made the piece?

If Disney were to tell animators to make a movie, and if lets say directed them by using blobs to show where things need to be drawn, Disney would still get to claim the copyright.

 

To abstract the idea even further, you aren't generating the art on the computer when doing 3D animation...instead you are just directing the computer what to do.

 

Specifically the line gets blurred on what should and what shouldn't be considered art attributable to a company.  Again, I would like to bring up my example from my previous post.  Would you consider it copyrightable if they trained it on drawings they did themselves?  Or lets look at Futurama, they essentially used 3d models to generate some of the 2d scenes.  At what stage does one lose their copyright of it?

 

[Brooksie359]

3 hours ago, LAwLz said:

Why is it good that AI generated art can't be copyrighted? 

One thing I could think of is that the AI potentially could spit out the same image more than once and then who would own the copyright? I mean technically the copyright should go to the AI itself and not someone who is simply utilizing the AI. 

[wanderingfool2]

2 minutes ago, Brooksie359 said:

One thing I could think of is that the AI potentially could spit out the same image more than once and then who would own the copyright? I mean technically the copyright should go to the AI itself and not someone who is simply utilizing the AI. 

A similar thing can be said about humans though.  If you ask it to draw a stickman drinking tea, there will be people who draw nearly the same thing (like not pixel perfect but close enough that they look the same at a glance).

 

Ignoring the fact that most AI models were trained on data that they later regurgitate out (as eventually I do think it will get to the point where it loosely gets the concepts of what a face should look like and doesn't do direct copies).  I think it should be held to similar types of copyright as it was a human; with the caveat that the copyright on it should be lesser given that it was lets say text prompt driven.

 

So for example, an image generated by putting in a full paragraph where it draws the scene should have more copyright protection than lets say "draw a stickman".  Even today with humans similar things exist, if you were to draw a stick man you could copyright it but really you won't be able to enforce it unless someone does a 1 to 1 copy of your work.

[Zodiark1593]

1 minute ago, Brooksie359 said:

One thing I could think of is that the AI potentially could spit out the same image more than once and then who would own the copyright? I mean technically the copyright should go to the AI itself and not someone who is simply utilizing the AI. 

Copyright protections can only apply to a "Legal Person".

 

Page 4 of the opinion from Naruto vs Slater (ie, the monkey selfie copyright dispute) does cover this.

 

"We must determine whether a monkey may sue humans,corporations, and companies for damages and injunctive relief arising from claims of copyright infringement. Our court’s precedent requires us to conclude that the monkey’s claim has standing under Article III of the United States Constitution.Nonetheless,we conclude that this monkey—and all animals, since they are not human—lacks statutory standing under the Copyright Act.1We therefore affirm the judgment of the district court. "

 

https://law.justia.com/cases/federal/appellate-courts/ca9/16-15469/16-15469-2018-04-23.html

[Holmes108]

3 minutes ago, Zodiark1593 said:

Copyright protections can only apply to a "Legal Person".

 

Currently. As always though, laws often have to change with the times. Whether or not this is one of those times is obviously the debate though.

[goodtofufriday]

1 hour ago, 05032-Mendicant-Bias said:

It takes more steps than you think to generate a decent looking image.

 

In this example I used OpenPose to get the pose from the left image, used ProtogenX34 guided by open pose to generate a 1024x1024 blurry mess with mishaped face and fingers, used GAN4X to upscale with a GAN network, then inpainted and regenerated individual details like hands, face, etc... All while refining positive and negative prompts.

It's not art skills because I can never draw anything even remotely close to the right image, no matter how much time I would choose to learn it. It is certanly creative skills to refine the prompt and get closer to what I had in mind. A D&D high elf mage character.

i get what youre saying about some skill needed here.

But legally its already been decided many times over,  even in the current case, that minor edits do not rise to the level of copywritable work when the initial image is largely indisginguishable.  

[jagdtigger]

16 minutes ago, Holmes108 said:

 

Currently. As always though, laws often have to change with the times. Whether or not this is one of those times is obviously the debate though.

Id say we shouldnt, for one the training sets are full of copyrighted images so its very doubtful if we can even classify the generated images as original "works".....

[SolarNova]

1 hour ago, goodtofufriday said:

If we had the food making machines from star trek, would you call someone who told a machine "make a michelin 5 star meal" a chef? Probably not. So why is someone who tells a machine "make this piece of art" an artist?

A chef uses tools ..kitchen appliances ..to make said meal. And is called a chef.

An artist uses tools...including digital ones ...to make art ...and is called an artist.

 

AI image generators are a tool, and the vast majority of results require tweaking in other programs to be properly presentable.

 

I can draw a scene using stick figures in a few seconds, it would look like a 5 year old drew it ....i wouldn't consider it art or myself an artist.

i could spend a week on a drawing digital or otherwise, and come out with something presentable and consider it artwork, and as such myself an artist for it.

 

The final product and objective matters. You cant just plug in a few key words, press generate and use the image as is and expect people to consider it 'art'. but you can use it as part of something that can be used to create a piece of art, or otherwise spend time tweaking the programs settings, the use of specific images, tuned keywords etc to come out with a results that requires minimal tweaking.

The fact a person used an image generator as part of the process shouldnt disqualify them from being considered an artist, the work art ,,or the copy-write protections afforded to it should they choose to pursue it.

A chef can use all many of modern technology to assist in the making of meals, many have a lot of automation involved, they are still a chef.

 

Much of the anti Ai image generator crowed honestly sound like butt hurt individuals worried about loosing money and or recognition as artist, or people with artistic skills, to those who would normally be unable to produce works for themselves in a convenient timely manner. I suppose its not abnormal, new tech that make things more accessible are usually decried by the people who have control over the 'old' original ways of production.

[Holmes108]

2 minutes ago, jagdtigger said:

Id say we shouldnt, for one the training sets are full of copyrighted images so its very doubtful if we can even classify the generated images as original "works".....

Humans are trained on copyrighted works and can ultimately make art that is sufficiently unique. I see no reason why AI won't be able too as well. But I also think that's only part of the puzzle, so not really sure where I land on this one.

[Thaldor]

32 minutes ago, wanderingfool2 said:

If Disney were to tell animators to make a movie, and if lets say directed them by using blobs to show where things need to be drawn, Disney would still get to claim the copyright.

 

To abstract the idea even further, you aren't generating the art on the computer when doing 3D animation...instead you are just directing the computer what to do.

 

Specifically the line gets blurred on what should and what shouldn't be considered art attributable to a company.  Again, I would like to bring up my example from my previous post.  Would you consider it copyrightable if they trained it on drawings they did themselves?  Or lets look at Futurama, they essentially used 3d models to generate some of the 2d scenes.  At what stage does one lose their copyright of it?

 

And?

 

That is exactly my question. Disney draw the Mickey Mouse and the other characters and example of the background. Why can Disney claim a copyright when he directs people to draw Mickey Mouse animation in his style but he cannot claim a copyright if he directs an AI to draw a Mickey Mouse animation in his style?

It isn't even a question about the directing method because you probably could direct AI with the same blobs you direct people and direct people with just by writing words on a computer screen as you do with certain AIs.

 

And just to lay it down, I am going into this with the same mentality as I go with the argument that the primary feature that makes a diamond valuable is that it was dug from ground and purity, size, color and other features are secondary which can only raise that initial value because we can create everyway possible superior diamonds in lab except they are worthless even when compared to significantly inferior natural one. The main difference between Jeff Koons directing a person to make a sculpture and Jeff Koons directing an AI to make a sculpture isn't Jeff Koons's artistic input to the sculpture, his handmark on the sculpture or whatever but that he steals the copyright. And somehow art world is absolutely fine with someone stealing the copyright from someone else but completely furious about someone claiming a copyright to a piece made by a machine. If we want to take it there I believe we could give the same contracts, NDAs and other paperwork and even a monthly wage to an AI as we do to a studio assistant and most likely that wouldn't change a thing.

 

As in we are cool with Jeff Koons basicly commissioning an painting from an artists and claiming it to be his. But God forbid if he was to commission an AI that doesn't even care what the piece actually is.

[Holmes108]

12 minutes ago, SolarNova said:

A chef uses tools ..kitchen appliances ..to make said meal. And is called a chef.

An artist uses tools...including digital ones ...to make art ...and is called an artist.

 

AI image generators are a tool, and the vast majority of results require tweaking in other programs to be properly presentable.

 

I can draw a scene using stick figures in a few seconds, it would look like a 5 year old drew it ....i wouldn't consider it art or myself an artist.

i could spend a week on a drawing digital or otherwise, and come out with something presentable and consider it artwork, and as such myself an artist for it.

 

The final product and objective matters. You cant just plug in a few key words, press generate and use the image as is and expect people to consider it 'art'. but you can use it as part of something that can be used to create a piece of art, or otherwise spend time tweaking the programs settings, the use of specific images, tuned keywords etc to come out with a results that requires minimal tweaking.

The fact a person used an image generator as part of the process shouldnt disqualify them from being considered an artist, the work art ,,or the copy-write protections afforded to it should they choose to pursue it.

A chef can use all many of modern technology to assist in the making of meals, many have a lot of automation involved, they are still a chef.

 

Much of the anti Ai image generator crowed honestly sound like butt hurt individuals worried about loosing money and or recognition as artist, or people with artistic skills, to those who would normally be unable to produce works for themselves in a convenient timely manner. I suppose its not abnormal, new tech that make things more accessible are usually decried by the people who have control over the 'old' original ways of production.

 

I agree strongly with everything you said. But the issue is still more complicated than a lot of that. That's why we need to do away with all the philosophical stuff. I don't care what "art" is, or an "artist". I don't care if the AI is sentient or not. I don't care whether we should or shouldn't give legal protections that were traditionally for humans to non-humans. 

 

To me, it really just comes down to the real life, practical, measurable impacts, both positive and negative that will happen to industry as a whole, depending on which way we decide to go with it. There are clear pitfalls either way, as discussed in this thread though. So I don't know which way will ultimately cause more harm or good. I don't believe anyone does.

 

Edit: Still a fascinating conversation though!

[Kisai]

1 hour ago, jagdtigger said:

Because its a generator and not a creator. It takes ppls work and use parts from them without any sort of modification and stitches them together......

 

That's describing how it works with a very naïve understanding of it. The "original" data is not in the model. Weights that may resemble the training data can sometimes result from over-fitting (seeing the same image too many times.) 

 

1 hour ago, jagdtigger said:

As for the topic its a good decision. Now the only thing we need is a ban on using copyrighted material without permission as training material.

That's unlikely going to happen. Closing off the way humans learn to machines just doesn't make sense, and is impractical.

 

Like there's no difference between showing the 512x512 jpeg of an image and pointing a 720p webcam at a LCD screen in this case.

 

Likewise, text (GPT3, and ChatGPT) and audio. Text is extremely hard to put a restriction on because of how much text is common, reused and referenced by others.  High quality audio is also super rare to have to be able to train from.

 

So a training corpus of "public domain" works would only result in the AI having the information from 75+ years ago, and we already have enough problems with racial and gender bias caused by training on limited sample sizes, we don't need to make it worse by only training on material from the time of extreme racism.

 

 

 

[goodtofufriday]

19 minutes ago, SolarNova said:

A chef uses tools ..kitchen appliances ..to make said meal. And is called a chef.

An artist uses tools...including digital ones ...to make art ...and is called an artist.

 

AI image generators are a tool, and the vast majority of results require tweaking in other programs to be properly presentable.

 

I can draw a scene using stick figures in a few seconds, it would look like a 5 year old drew it ....i wouldn't consider it art or myself an artist.

i could spend a week on a drawing digital or otherwise, and come out with something presentable and consider it artwork, and as such myself an artist for it.

 

The final product and objective matters. You cant just plug in a few key words, press generate and use the image as is and expect people to consider it 'art'. but you can use it as part of something that can be used to create a piece of art, or otherwise spend time tweaking the programs settings, the use of specific images, tuned keywords etc to come out with a results that requires minimal tweaking.

The fact a person used an image generator as part of the process shouldnt disqualify them from being considered an artist, the work art ,,or the copy-write protections afforded to it should they choose to pursue it.

A chef can use all many of modern technology to assist in the making of meals, many have a lot of automation involved, they are still a chef.

 

Much of the anti Ai image generator crowed honestly sound like butt hurt individuals worried about loosing money and or recognition as artist, or people with artistic skills, to those who would normally be unable to produce works for themselves in a convenient timely manner. I suppose its not abnormal, new tech that make things more accessible are usually decried by the people who have control over the 'old' original ways of production.

A chef and real artist physically manipulate the parts of what they are creating and are they themselves changing the shape and form what they interact with.

 

Ai as a tool removes that process. That process is what humans consider to be what is copywritable. 

 

We can also look at relatable cases of Elonor the car. its a shelby gt mustang with some slight modifications. The case for getting elonor a copyright failed as it was determined not to be distinct enough from the original work of the shelby gt. The modification took real work, skill, and imagination to complete. But it was still not enough to grant copyright.
Ai art would be the same. You have people making modifications to the original generated image, but this is not enough to rise to the level of copyright. So neither the generate image nor the touched up ones qualify for copyright. 

 

Im an IT admin and do work on the hardware side of things. I dont have a horse in this race as until Ai develops to a point where they are robots im not as risk of losing my livelyhood. So i feel my opnion here is from an unbiased view.